Nala Krisnanda

I have a lot of learning interest, more specifically in data, matemathics, and engineering. Welcome!

A data practioner with 3 years of experience in Indonesian tech companies. Currently a Statistics with Data Science MSc student at The University of Edinburgh, aiming to obtain expertise in analytics, optimization, and causal inference.

Career

Data Analyst

Traveloka → Nov 2021 - Aug 2023

I was part of the shared services team, specifically in payment fraud. My role primarily revolved around enhancing the fraudulent activity monitoring such as chargeback, PayLater fraud, coupon abuse, etc. Additionally, I also provided the necessary data and analysis to the stakeholders to aid their investigative efforts.

Being part of multiple enhancements to our fraud detectors has provided me hands-on exposure to ML in practical setting, allowing me to implement some models including decision tree and logistic regression for detecting fraud patterns. Moreover, I was actively exploring network approach for fraud detection which gave me the opportunity to deepen my understanding and utilize some of the graph features.

Achievement:

  • Responsible in developing an end-to-end fraud monitoring system which inolved tasks such as designing tables, maintaining data pipelines, and creating dashboards. This achievement has enabled the company to more efficiently track overall fraud figures across different verticals.

Data Analyst

Tokopedia → Sep 2020 - Nov 2021

Worked with TopAds, which is a product that allow seller to promote their products in Tokopedia platform. My role was mostly optimizing the ads ranking in the search result page.

This experience helped me learn about ML, like regression and learning-to-rank models. I also got to explore some optimization methods, such as genetic algorithms and NSGA. Lastly, this role has given me an in-depth understanding about experimentation (A/B testing) design and analysis.

Achievement:

  • Improve Tokopedia ads ranking strategy (in search result, year 2021) by using learn-to-rank and regression machine learning method. This is successfully increase the conversion rate by 2.14% and TopAds revenue by 3.22%
  • Built a centralized ads scoring package in Python.
  • Constructed a centralized post-experimentation analysis package in Python with several features e.g. hypothesis testing with p-value correction, estimating variance with delta method, and variance reduction with cuped.

Prior to 2020

Education

Statistics with Data Science

University of Edinburgh → 2023 intake period

Computer Engineering

Universitas Indonesia → Graudated in 2020

Developed a CNN model to detect driver drowsiness level using EEG, heart rate, and eyelid indicators, then integrated it into an Android app with Tensorflow Lite.
Repo Demo

Misc